Test Data Generation with A Hybrid Genetic Tabu Search Algorithm for Decision Coverage Criteria

In order to improve the efficiency of test data generation during the software test, many studies have been focusing on the automation of test data generation with metaheuristic techniques such as genetic algorithm and tabu search etc. This paper proposes a hybrid algorithm in combination with genetic algorithm with tabu search to generate the test data based on decision coverage criteria. The hybrid algorithm may improve the global search ability of optimizing the solution of tabu search and local search ability of genetic algorithm; at meanwhile, an approach was described in detail to insert fitness function as stubs in source code, control the prematurity during search iteration, avoid falling into the local optimum and speed up the convergence speed. By applying the hybrid algorithm to the instance program under test, the final result shows that both the decision coverage and the calculation speed have been improved.